Quote:
Originally Posted by mathjak107
these are not computer simulations , the safe withdrawal numbers are based on actual worst case real returns over every single 30 year period.
the events that caused them do not matter only the results which are if you get less than a 1% average real return per year over any 15 year time frame you will most likely have to take a pay cut from whatever your safe withdrawal rate is and you will have little cushion left for the unexpected expenses..
it is not always 4% . the number is dependent on age and number of years you want to plan out to and allocations.
sure ,you can pull 4 ,5,6% from any amount of savings regardless of allocations but what amount of risk of being one of the ones who failed in the past do you want to chance?
that is all retirement planning with these studies tells you. if you want the maximum cushion for all those unknown expenses you speak of I certainly would run with something that failed less in the past.
as I said trying to pull 4 or 5% from bonds only, has failed over and over through out so many time frames.
betting your retirement that you will be the case where it was okay may be foolish.
again these were never meant to be life long rules. they are planning tools to get you in a relatively safe ball park for your savings amount and the amount you want to draw.
it is up to you to season to taste.
|
Of course they're simulations - because no one in the studies actually ever did it.
And you kind of got me interested. So I went and looked at what data these computer simulations are using. Apparently only the 10 year note (which has been around for a long time - but which was for the longest time something only available to institutional size investors). *And* the return on the SP500. Now the SP500 was only introduced in 1957 (its first predecessor - which didn't have 500 stocks - was introduced in 1923).** And it would have been impractical/impossible for individual investors to own the SP500 before 1976 - when Vanguard introduced the first SP500 mutual "index" fund (the SP500 ETF - SPY - was only introduced in 1993).
IOW - based on the available data - assuming you're comparing apples with apples - one can only make computations from 1957 (introduction of the SP500) to 1983 (the last full 30 year period ended in 2013). Which is only 27 time periods. Not exactly an overwhelming number of time periods.
FWIW - this is why I always hated "black box" trading systems. When someone (especially me!) is relying on data to do something - I want to know what the data is.
Also - Pfau is making certain assumptions about *future* market returns in his projections:
As a result [of current financial asset yields/prices], Pfau estimates a fifty-fifty portfolio of stocks and bonds is likely to deliver a long-run annual average return after inflation of just 2.2%, less than half thAs a result, Pfau estimates a fifty-fifty portfolio of stocks and bonds is likely to deliver a long-run annual average return after inflation of just 2.2%, less than half the historical rate.
And kind of assuming facts not in evidence:
Of course they're simulations - because no one actually ever did it.
And you kind of got me interested. So I went and looked at what data these computer simulations are using. Apparently only the 10 year note (which has been around for a long time - but which was for the longest time something only available to institutional size investors).* *And* the return on the SP500. Now the SP500 was only introduced in 1957 (its first predecessor - which didn't have 500 stocks - was introduced in 1923). And it would have been impractical/impossible for individual investors to own the SP500 before 1976 - when Vanguard introduced the first SP500 mutual "index" fund (the SP500 ETF - SPY - was only introduced in 1993).
IOW - based on the available data - one can only make computations from 1957 (introduction of the SP500) to 1983 (the last full 30 year period ended in 2013). Which is only 27 time periods. Not exactly an overwhelming number of time periods. Of course - if you trust the "blended" SP data - you can go back to 1928 - which gives you another 29 time periods. For a total of 56 annual time periods. I don't think that's enough data to base my retirement on.
FWIW - this is why I always hated "black box" trading systems. When someone (especially me!) is relying on data to do something - I want to know what the data is.
Also FWIW - Pfau is making certain assumptions about *future* market returns in his projections:
As a result [of current financial asset yields/prices], Pfau estimates a fifty-fifty portfolio of stocks and bonds is likely to deliver a long-run annual average return after inflation of just 2.2%, less than half the historical rate.
And assumptions about asset price correlations:
Even given the same long-run returns, a less volatile portfolio will tend to last longer. A portfolio that's not just large-cap stocks plus bonds, but is instead spread out to include small-company stocks, foreign stocks, and other asset classes, Pfau says, lets the safe drawdown from $1 million in savings rise to an initial $35,000 from $30,000.
Forget the 4% withdrawal rule - Feb. 26, 2014
Small cap stocks? The Russell 2000 first came out (as an index) in 1984. I buy/hold/trade the ETF (IWM) - which has only been around since 2000. We forget how new some of this stuff is. I sure don't expect the IWM to go up when the SP500 is going down. Indeed - in recent years - more and more asset classes that were traditionally non-correlated have become increasingly correlated:
RICH BERNSTEIN: 3 Things You Should Know About Financial Asset Correlation - Business Insider
The primary substantially non-correlated equities asset today is treasuries. OTOH - until recently - these assets tended to be correlated - a relationship that might reappear if interest rates were to rise a fair amount
Finally - a lot of these studies/this data - at least on the surface (perhaps it's all buried in footnotes) ignore a lot of important things. Like what do people do with interest/dividends? Spend or reinvest? If the latter - at what prices? What tax rates do people pay when they receive interest/dividends or sell assets (IRAs were first introduced in 1974). People have been known to harvest both capital gains and capital losses at times
. What about commissions? Do these studies assume "frictionless" investing (minus taxes and fees)? If so - they're not realistic. Robyn
*Using only the 10 year note as the "bond proxy" is problematic for several reasons IMO. First - individual investors usually don't buy 10 year notes (and they're much much easier to buy today than in the past - and we're not talking ancient past either). Most non-institutional investors bought CDs - and usually shorter term ones. The yield was often better - but buying short term CDs leaves one unduly exposed to reinvestment risk. Individual investors with more money tended to buy munis - which have almost always offered better/much better after tax yields than the 10 year.
**This is the best history of the SP500 I could find:
History and Structure of the Standard & Poor's 500
The new SP500 data was tacked on to the data from an older different index so a body of historical data could be created. Whether this is good or bad - I don't know. All I know is you're dealing with data for 2 different indices.